Ceph configuration issues - Health problems

[tangerine]

Hi All,

I've done my best to read documentation and research the forums / web. I keep having these health errors pop up. I've tried a ton of different PG configurations including 128, 1024, 2000, etc. I can't seem to nail this setup - I've tried using calculators as well.

I have 3 nodes, 8 disks each, each disk is 2TB (1.8TB). Any help and I'd be grateful.

 

[tangerine]

To add - in all scenarios I have OSD's going up and down (mainly on one specific node).

 

[tangerine]

Figured it out.

I need to review my switch config. Server is setup on a bonded interface running LACP. When I shutdown the port on one of two switches it resolves the OSD flapping and Ceph errors.